Cell culture and reagents

The induction and maintenance of iPSC-derived neural crest cells (iNCCs) and MSCs (iMSCs) were conducted as previously described, with minor modifications [16, 17]. Briefly, iPSCs (1231A3) [27] were maintained in StemFit AK03N (Ajinomoto, Japan) supplemented with 10 µM Y27632 (Wako Pure Chemical, Japan) in dishes pre-coated with iMatrix-511 (nippi, Japan). iNCCs were induced in StemFit Basic03 (equivalent to AK03N without basic fibroblast growth factor (FGF2), Ajinomoto, Tokyo, Japan) with 10 µM SB-431542 (Selleck Chemicals, USA) and 1 µM CHIR99021 (Axon Medchem, Netherlands) for 10 d. iNCCs were maintained in StemFit Basic03 with 10 µM SB-431542 and 20 ng/mL recombinant human FGF2 (WAKO) and recombinant human EGF (R&D Systems, USA) onto plates pre-coated with fibronectin (Merck, USA). iMSCs were induced and maintained in PRIME-XV MSC Expansion XSFM (Fujifilm Irvine Scientific, USA) onto the fibronectin-coated plates. FOP-iMSCs and resFOP-iMSCs were generated as previously described [11, 16] and cultured in αMEM (Invitrogen) supplemented with 5 ng/ml FGF2 (WAKO), 10% fetal bovine serum (FBS; Nichirei, Inc., Japan), and 0.5% penicillin and streptomycin (Invitrogen). The reagents are listed in Additional file 1: Table S1.

Construction of stably expressing cells

Samples of pPB-CAG-ACVR2B-Fc-His-Puro (5 µg, ACVR2B-Fc) and pPV-EF1a-EGFP-IRES-Puro (5 µg, Control) plasmids were mixed with 1 µg of pHL-EF1a-hcPBase-A plasmid (PBase) in Opti-MEM (Gibco, USA). For IVIS imaging assay, 5 µg of pPV-EF1a-Luciferase-BleoR and 1 µg of PBase plasmids were mixed with 5 µg of pPB-CAG-ACVR2B-Fc-His-Puro and 5 µg of pPV-EF1a-EGFP-IRES-Puro. A total of 100 µl of the DNA and iMSC mixture was obtained. According to the manufacturer's instructions, electroporation was performed using a NEPA21 electroporator (Nepagene, Japan). Stably expressing cells were selected using 1 µg/ml puromycin or 1 µg/ml puromycin (Gibco) with 400 µg/ml zeocin (InvivoGen) for 7 d. FACS analysis for positive MSC markers was performed as previously described [24].

Quantitative RT-PCR

Total RNA was isolated using an RNeasy Mini Kit (Qiagen, USA) and treated with a DNase-one Kit (Qiagen) to remove any genomic DNA, according to the manufacturer’s instructions. Thereafter, 300 ng RNA was used to perform first-strand cDNA synthesis using SuperScript™ III Reverse Transcriptase (Invitrogen, USA). Quantitative PCR was performed in triplicates using Thunderbird Next SYBR qPCR Mix (TOYOBO, Osaka, Japan) and QuantStudio 3 RealTime PCR System (Applied Biosystems, Forester City, CA, USA). The primer sequences are listed in Additional file 1: Table S2.

Western blotting analysis

To assess any phosphorylated proteins, iMSCControl and iMSCACVR2B-Fc cells were starved using 10% XSFM + 90% αMEM with or without 1 µM FK506 for 24 h and treated with the indicated ligands for 1 h. resFOP-iMSC and FOP-iMSC cells were serum-starved using αMEM for 3 h and treated with conditioned medium from iMSCControl or iMSCACVR2B-Fc (supplemented with the indicated ligands) for 1 h. Protein samples were harvested using RIPA buffer (Wako Fujifilm) supplemented with Halt Protease and Phosphatase Inhibitor Cocktail (Thermo Fisher Scientific). After sonication (Bioruptor), the protein concentrations were detected using Pierce BCA Protein assay reagent (Thermo Fisher Scientific). Next, 10 µg protein was mixed with 6X sample loading buffer (Tokyo Chemical Industry) and incubated at 95 ℃ for 10 min. SDS-PAGE was performed using e-PAGE 10% gel with WSE-1150 Page Run Ace (ATTO, Japan). Blotting was performed using iBlot™ 2 Transfer Stacks (PVDF) with an iBlot™ 2 Gel Transfer Device (Invitrogen). Non-specific antigens were blocked with Blocking One-P (Nacalai Tesque, Japan) for phosphorylated protein analysis and 5% skim milk in TBST for total protein analysis. Phosphorylated antibodies were diluted using Can Get Signal Immunostain Solution B (TOYOBO). All the other antibodies were diluted in 5% skim milk in TBST. Signals were detected with Amersham ECL Prime Western Blotting Detection Reagent (Cytiva) and visualized using an Amersham-ImageQuant-800-biomolecular-imager (Cytiva). Conjugated antibodies on blotting membrane were removed using WB Stripping Solution (Nacalai Tesque). Washed blotting membranes were used for a second antigen–antibody reaction. The antibodies used in western blotting are listed in Additional file 1: Table S3.

ELISA

Cell culture supernatants were centrifuged at 10,000 × g and 4 ℃ for 10 min. His-tagged ACVR2B-Fc proteins were detected using a His-Tag ELISA Detection Kit (L00436, Genscript). Blood samples of mice were left at room temperature (RT, 20–25 ℃) for 30 min until clot formation. The mouse serum was harvested after centrifugation for 15 min at 2,000 × g. Activin-A and Fc fragment in serum were detected using Human/Mouse/Rat Activin-A Quantikine ELISA Kit (DAC00B, R&D Systems) and Human Fcγ (Fc Fragment of IgG) ELISA Kit (E-EL-H1745, Elabscience). The absorbances at 450 nM were read using an EnVision Multilabel Reader (PerkinElmer).

Luciferase assay

Cells were seeded at 1 × 105 per well with 500 μl XSFM medium onto a 24-well plate. The following day, the cells were transiently transfected with plasmids as pGL3-BRE-Luc [28]: pRL-CMV-renilla = 10:1 or pENTR(CAGA)9-Luc2 [11]:pRL-CMV-renilla = 10:1. pRL-CMV-renilla was used as the internal control. FuGene® HD Transfection Reagent (Promega Corporation, USA) was used for transfection for 24 h at a 3:1 ratio to DNA, according to the manufacturer’s instructions. The iMSCControl and iMSCACVR2B-Fc cells were stimulated with indicated ligands without a medium change. For resFOP-iMSCs and FOP-iMSCs, the medium was replaced with conditioned medium from iMSCControl or iMSCACVR2B-Fc after 24 h of transfection and supplemented with the indicated ligands. After incubation for 16 h, luciferase activity was measured using a dual luciferase reporter assay system (Promega) on an EnVision® Multilabel Reader (PerkinElmer).

FOP mouse model

The cardiotoxin-induced ACVR1R206H conditional transgenic mouse model was described previously [20]. Briefly, female mice aged 16 − 21 weeks were administrated with drinking water supplemented with 2 mg/ml doxycycline hyclate (Dox, Sigma Aldrich) and 10 mg/ml sucrose (Nacalai Tesque) to initiate the ACVR1R206H expression. The right gastrocnemius muscle injury was induced via cardiotoxin (CTX, 9.1 μg/mouse; latoxan) injection. Mice were randomly allocated into each group (cage) using a random-number table. For the primary HO model, four injections of 1.5 × 106 cells in 100 µl αMEM were locally or intraperitoneally administered every 5 d using a 27G syringe. For the recurrent HO model, the HO tissues were surgically resected as much as possible under isoflurane anesthesia after 2 weeks of CTX-initiated muscle injury. The wounds were sutured steadily using 5–0 stitches (Bear Medic Corporation, Japan). Three injections of 3 × 106 cells were administered intraperitoneally every 4 d, starting the day after surgery. The mice that died before the endpoint of the experiments were excluded from the analysis. The mouse blood sampling was performed through cardiac puncture under isoflurane anesthesia. Subsequently, specimens were harvested after euthanasia via carbon dioxide inhalation. Our manuscript reporting adheres to the ARRIVE guidelines in accordance with BioMed Central editorial policies.

In vivo imaging

After injecting 150 mg/kg d-luciferin substrate (Summit Pharmaceutical International Corporation, XLF-1) diluted in PBS for 15 min, mice were anesthetized using isoflurane (Pfizer, USA) and imaged using an IVIS Lumina series III (PerkinElmer). The region of interest was selected to cover all signals, and luminescent intensity was measured using Living Image Software (Caliper Life Sciences, USA).

Rotarod and treadmill test

After adaptive training, mice were loaded onto a rotarod station (ENV-574 M, Med Associates, USA) and Rodent Treadmill (Ugo Basile SRL, Italy). The rotation speed was set from 2 to 20 rpm and maintained at 20 rpm until the mice fell. The riding times were recorded. The treadmill movement test (+ 5º inclination) started at a 5 m/min velocity, gradually increasing to a maximum of 20 m/min (shock: 1 Hz, shock intensity 0.2 mA). Exhaustion was defined as the mice being shocked 15 times or not returning to the treadmill and staying on the shock grid for 5 s. Running distances were recorded.

X-ray and micro-computed tomography (μCT) imaging

A mixture of medetomidine, midazolam, and butorphanol was injected intraperitoneally to anesthetize the mice. X-ray images were acquired using an AB-35 system (Acrobio, Japan). The μCT images were scanned using inspeXio SMX100CT (Shimadzu, Kyoto, Japan) and analyzed using FCS-Bon64 (Ratoc System Engineering, Tokyo, Japan) software.

Histochemical and IHC analyses

Tissue samples from the mice were fixed with 4% paraformaldehyde for 24 h, decalcified in 12% EDTA for 10 d, embedded in paraffin, sectioned, deparaffinized and stained with hematoxylin and eosin (H&E) and Alcian blue. For IHC staining, antigen retrieval was performed via incubation with Liberate Antibody Binding Solution (Polysciences, USA) for 20 min at RT. Non-specific antigens were blocked using Blocking One (Nacalai Tesque) for 1 h at RT. For antibodies generated from the mouse host, tissues were blocked with ReadyProbes™ Mouse on Mouse IgG Blocking Solution (Invitrogen) for 60 min at RT. Antibodies (diluted in Can Get Signal Immunostain Solution B, TOYOBO) were incubated overnight at 4℃. After rinsing in PBST buffer, samples were incubated with Alexa Fluor secondary antibodies diluted in Can Get Signal Immunostain Solution B for 1 h at RT. A mounting medium with DAPI (Vector Laboratories) was used to counterstain the nuclei. Images were captured using a BZ-X810 microscopy (Keyence, Japan). The antibodies for IHC are listed in Additional file 1: Table S3.

Statistical analysis

The statistical significance of all experiments was calculated using unpaired t-tests, one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons tests, two-way ANOVA with Šídák’s multiple comparisons tests, or two-way ANOVA with Tukey’s multiple comparisons tests. The tests used are indicated in each figure legend. All statistical analyses were performed using GraphPad Prism 9 (GraphPad Software). P values < 0.05 were considered statistically significant. Significance levels are: *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.